SPACEL.Scube.gpr.GPRmodel.train

GPRmodel.train(lr=1, training_iter=500, save_model=True, save_pred=False, cal_bf=False, optim_l=False, optimize_method='Adam')

Training GPR model

Training GPR model.

Parameters:

  • lr – The learning rate used in the training process.

  • training_iter – The number of iterations for the training.

  • save_model – A boolean value indicating whether to save the trained model. The default value is True.

  • save_pred – A boolean value indicating whether to save the prediction results.

  • cal_bf – A boolean value indicating whether to calculate the BF (Bayes factor) value.

  • optim_l – A boolean value indicating whether to optimize the initial length scale value.

  • optimize_method – The optimization method used in the training. It must be one of ‘Adam’ and ‘LBGFS’. By default, it is set to ‘Adam’.

Returns:

None